Sorry, lots of things to do these last weeks. Marton Balint via ffmpeg-devel (HE12025-08-26): > I just thought that for the typical case we can spare an strlen() call this > way.
Except if the format is pathologically long, which is not a case we want to optimize for, I suspect the branching is more expensive than the counting. > > I think a lower margin should be ok, >>4 or >>5. > I'd rather keep it >>8, because strftime format strings can also have width > specifiers, so something like "%80c" should work. Obviously this is always > going to be a heuristics, but I wanted to be careful and reject only the > very unlikely/insane format strings... I am worried that this large value can cause bprint to go beyond the built-in buffer and allocate more memory just to be sure. But we are in a pathological case anyway, so go ahead. > There is one, but OK, I will make the existing comment a bit more verbose. Thanks. Regards, -- Nicolas George _______________________________________________ ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org To unsubscribe send an email to ffmpeg-devel-le...@ffmpeg.org
